Economic Review: July 2015

• The Quarterly National Accounts (QNA) indicated that UK GDP grew by 0.4% in Q1 2015, 0.1 percentage points higher than the previous estimate. Growth in 2014 as a whole was also revised up, from 2.8% to 3.0%.

• Comparing the three months to April 2015 with the same three months of the previous year, average weekly earnings ‘total pay’ increased by 2.7% - slightly stronger than the median growth rate of pay over the same period. This may indicate an easing of downwards pressure on earnings growth from compositional effects.

• In 2013/14 the net impact of all tax and benefits measures was for income to be redistributed from high income households towards poorer households, resulting in a more equitable income distribution. Real household disposable income fell in both the first (-0.5%) and fifth income quintiles (-1.6%) in 2013/14, but increased for the second, third and fourth quintiles.

• UK house prices rose by 5.5% in the year to April 2015, and by 10% over 2014 as a whole. However, this masks a wide range of growth rates among the local authorities of England and Wales.

The Quarterly National Accounts (QNA) indicated that GDP rose by 0.4% in Q1 2015; slightly higher than previously estimated, but slower than in recent quarters. There were also small upward revisions to growth in the second, third and fourth quarters of 2014. These were mainly a result of methodological changes to construction output price indices, which affected construction output and gross fixed capital formation (GFCF).

This edition of the Review examines the recent increase in whole economy earnings growth alongside evidence on the median growth rate of earnings for those in continuous employment. The recent strength of both measures suggests a rise in wage pressures in the UK labour market. The closing of the gap between average weekly earnings growth and the median growth rate of earnings may reflect some easing of the compositional effects which have hampered wage growth in recent periods. The recent rise in wage growth has been accompanied by a rise in productivity, on both a quarterly and annual basis.

Despite recent improvements in household earnings, real household disposable income was broadly unchanged on the quarter, partly as a result of increased taxes on income and wealth and decreased social benefits. This Review considers the impact of these policies on real household disposable income across the income distribution. It highlights recent ONS analysis which suggests the net effect of taxes and benefits is to redistribute income from households at the top of the income distribution towards households with lower incomes. Real household disposable income fell in both the first (-0.5%) and fifth income quintiles (-1.6%) in 2013/14, but increased for the second, third and fourth quintiles.

Average house prices in the UK rose by 5.5% in the year to April 2015, and by 10% over the 2014 calendar year. However, this masks a wide range of growth rates, with 10% of local authorities in England and Wales experiencing house prices rise by over 13.5%, while 10% experienced a rise of less than 2.1%. This Review finds that the highest value regions have experienced the highest rates of growth since the downturn and examines the distribution of house price to earnings ratios at the local authority level.

The Quarterly National Accounts (QNA) indicated that the UK economy grew by 0.4% in the first quarter of 2015: an upward revision of 0.1 percentage points from the previous estimate. Although weaker than the average quarterly growth rate since the start of the economic recovery, this was the ninth quarter of GDP growth and continues the upward trend in output which started in 2013.

The QNA also contained new information on broader indicators of economic well-being for Q1 2015, some of which are presented alongside headline GDP in Figure 1. The level of GDP per capita – which captures the effects of growth in the economy as well as growth of the population – rose by 0.2% compared to the previous quarter, but is still 0.6% below its pre-economic downturn peak. Real Net National Disposable Income (RNNDI) per capita – which accounts for capital consumption, and measures the income available to residents of a country – remained 3.8% below its pre-downturn peak in Q1 2015, but rose by 1.6% on the quarter. Real Household Disposable Income (RHDI) per capita, which has been relatively stable in recent years, has remained flat on the quarter but has increased by 3.9% from the same quarter of the previous year.

Figure 1: GDP, GDP per capita, real net national disposable income per capita and real household disposable income per capita, chain-volume measure, seasonally adjusted

Source: Office for National Statistics

Notes:

1. Quarter 1 (i.e. Q1) refers to January to March, quarter 2 (i.e. Q2) refers to April to June, quarter 3 (i.e. Q3) refers to July to September, and quarter 4 (i.e. Q4) refers to October to December.

Alongside the higher estimate of GDP growth in Q1 2015, the QNA also contained small upward revisions to quarterly GDP growth through 2014. Output growth is now estimated to have been 0.1 percentage points stronger in each of the second and third quarters of 2014, and 0.2 percentage points higher in Q4 2014. As a consequence, growth in 2014 has been revised up from 2.8% to 3.0%. Taking the year as a whole, household spending (which added 1.6 percentage points to expenditure growth) and gross capital formation (1.6 percentage points) were the main drivers of GDP. Export growth added just 0.1 percentage points over the same period, which was more than offset by the growth of imports, resulting in net trade making a substantial negative contribution to expenditure growth in 2014.

Following these revisions, output is now estimated to have been 2.9% higher in Q1 2015 than the same period a year ago – 0.5 percentage points higher than the previous estimate. Figure 2 decomposes these revisions to GDP growth to show the contributions of the expenditure (LHS) and output (RHS) components to the revision. The main driver of the revisions was the change in methodology for construction price and cost indices. This led to a 0.3 percentage point upward revision to Gross Fixed Capital Formation (GFCF) growth in the year to Q1 2015, affecting the expenditure measure of GDP. Gross value added was also affected by a 0.3 percentage point upward revision to construction output – the effects of both revisions are shown in Figure 2. ONS have published further details regarding the new construction price indices, as well as two articles that explain how these new prices have affected the volume measures of construction output (214.3 Kb Pdf) and investment spending (196.6 Kb Pdf) . Higher growth in household consumption and services output also augmented the whole economy revisions on the expenditure and output side respectively.

Figure 2: Revisions to expenditure and output measures of GDP and GVA growth respectively, quarter on same quarter a year ago, chained-volume measure, seasonally adjusted, percentage points

Source: Office for National Statistics

Notes:

1. 'Other' includes changes in inventories, 'acquisitions less disposals of valuables' and the statistical discrepancy.
2. Quarter 1 (i.e. Q1) refers to January to March, quarter 2 (i.e. Q2) refers to April to June, quarter 3 (i.e. Q3) refers to July to September, and quarter 4 (i.e. Q4) refers to October to December.

Despite the relative strength of the economic recovery since 2013, nominal earnings growth has been weak, and after their longest fall on record, real earnings only recently returned to growth. Past editions of the Economic Review, the Bank of England’s quarterly Inflation Report and work by Bank staff have highlighted that recent movements in earnings have been influenced by the changing composition of the work force. As earnings growth returns, the source of these increases is a key consideration. If earnings growth is a result of general inflationary wage pressure, these effects are likely to lead to higher average unit labour costs. If, however, earnings growth is a consequence of a shift in composition towards more highly-paid employment, the impact on within-industry unit labour costs may be more modest. Accounting for the impact of these composition effects is therefore key to assessing underlying wage inflation in the economy and evaluating whether the current uptick in wage growth is the start of a longer-run trend.

To give an indication of underlying wage growth (i.e. excluding composition effects), Figure 3 shows the median growth in earnings of those remaining in the same job from the Labour Force Survey. This measure tracks employees over a year, measures their individual earnings growth rate, and reports the median of these individual growth rates. To limit possible composition effects, such as people entering or leaving the workforce, or changing jobs, this measure only includes those who have been with the same employer (or self-employment) for one year or more.

Figure 3: Annual growth in average weekly earnings ‘total pay’, median growth in earnings for those in the same job for over one year, and Consumer Price Inflation, 3 month averages

Source: Labour Force Survey - Office for National Statistics

Over the long-term, Average Weekly Earnings (AWE) ‘total pay’ growth and median growth in earnings track each other fairly closely, reflecting the gradual nature of changes in workforce composition, set against differences in the mean and median concepts. Since the economic downturn, however, there have been two periods of clear divergence. In 2008 and 2009, much of the difference between the series is accounted for by changes in the timing of bonus payments related to changes in tax rates. However, some of the difference may be due to increased flows between employment and unemployment during and immediately following the downturn. These flows may have had a negative impact on AWE if those hired from unemployment received, on average, a lower wage than those currently employed. These trends, together with other factors, supported the continuously employed LFS measure to a greater degree than the all-employees AWE series.

The second period – in 2014 – coincides with a shift in the composition of the workforce. In this period, aggregate employment shifted towards lower-skilled occupations, as more entrants to the labour market took on lower-paid posts. This affected the all-employee, AWE measure, to a greater degree than the continuously employed group – among whom earnings growth was relatively robust. The closing of the gap in recent months may provide some evidence that these compositional effects are starting to wane, with implications for wage growth in coming months.

Changes in the composition of the workforce – and in particular the allocation of labour among different industries – may also affect labour productivity in the long run. Typically wages will be higher in industries in which the level of productivity is higher. Workers that are more productive will generate more output and as a result receive higher earnings. Therefore, the impact that industry composition has on average earnings may be associated with a similar composition effect on productivity.

One way to assess how the composition of the workforce may affect productivity, that goes wider than the industries workers are employed in, is to examine quality adjusted labour input. This is a measure which weights hours worked according to their total share of labour income. The “quality of labour” fell slightly during 2014, providing further evidence of a negative workforce composition effect over this year.

These recent negative effects notwithstanding, higher earnings growth in recent months has been accompanied by stronger productivity growth. Output per hour – a key measure of labour productivity – grew by 0.3% in the first quarter of 2015, and was 1.3% higher in Q1 2015 than in the same period a year earlier, the strongest annual growth in this measure since the start of 2012.

Earnings from employment – whether measured by the Average Weekly Earnings, the Annual Survey of Hours and Earnings or the Labour Force Survey – only capture one component of pre-tax household income. A more comprehensive picture – including information about other income sources and the impact of the tax and benefits system – is presented in the ONS’ annual ‘Effects of Taxes and Benefits’ publication. This article sets out households’ original income, the impact of benefits in cash and in kind, and the impact of direct and indirect taxes on households at different points in the income distribution.

Figure 4 shows that the impact of all tax and benefits measures is for income to be redistributed from high income households towards poorer households, resulting in a more equitable income distribution. In 2013/14, the poorest fifth of households (those in the bottom quintile group) had a weekly income of £106 before taxes and benefits. After redistribution, this rose by £191, or 180%, to £297. Income in the second and third quintiles also grew after the effects of taxes and benefits, but by smaller amounts, while the incomes of the richest two quintiles fell after the introduction of taxes and benefits, by £86 (11.0%) and £398 (25.7%) respectively.

Figure 5 decomposes the difference between original and final income into benefits in kind, cash benefits, as well as direct and indirect taxes. This shows that the richest fifth of households pay £398 more in taxes than they receive in benefits each week, while the second quintile receives more in benefits than the first quintile, more than offset by a greater tax contribution.

How do the benefits that households receive vary across the income distribution? The largest contribution to cash benefits for all income quintiles comes from state pensions; in the first quintile these contribute 35% of the £7,394 the average household received as direct benefits in cash in 2013/14. This proportion rises gradually to 67% for the richest fifth of households, who received £2,947 as benefits in cash on average. Benefits in kind received by the poorest fifth of households were also greater than those received by the wealthiest, with the largest difference coming from education, reflecting the fact that poorer households are more likely to enrol their children in state education.

Figure 5 also highlights the important role that direct taxes play in the redistribution of income. The richest fifth of households pay £377 in direct taxes each week, over 15 times that of the poorest households – much of which is accounted for by differences in income taxes. The burden of indirect taxes is also larger in absolute terms for higher income households – reflecting their higher level of consumption – although indirect taxes account for a larger fraction of disposable income for poorer households. Indicative of the different spending patterns of the population, the richest fifth paid over 8 times more in air passenger duty than the poorest fifth in 2013/14.

Figure 5: Summary of the effects of taxes and benefits on ALL households, financial year ending 2014

Source: Office for National Statistics

While the bottom income quintile has typically received the largest net benefit from the tax and benefit system, Figure 6 shows that the long run average disposable income growth of this group underperformed compared to all other household types. Since 1977, the poorest quintile has seen their real equivalised disposable income rise by 77% while the richest quintile had a corresponding increase of 139%. However, these differences have been fairly stable in recent years, with all quintiles experiencing broadly similar disposable income growth. Indeed, the majority of the disparity in incomes between the richest fifth and poorest fifth opened up in the late 1980s and early 1990s, and has held relatively constant since then.

The different quintiles have also varied in their experiences over the economic downturn and subsequent recovery. In the 2008/09 and 2009/10 financial years, real household disposable income in the bottom quintile rose by 2.1% and 4.3% respectively, reflecting lower original income but higher cash benefits and lower direct taxes. In 2008/09, real household income fell for all other quintiles, returning to growth for all but the top quintile the following year. Over the course of the recovery, income growth has been weak for all groups. On average, real household income fell by 0.5% per year between 2007/08 and 2013/14, with the top 20% of households by income experiencing larger reductions. Over the most recent year, real household income fell by 0.5% for the first income quintile, and by 1.6% for the fifth income quintile, but increased for all other quintiles.

Figure 6: Growth in mean equivalised real household disposable income by quintile group, 1977-2013/14

Source: Office for National Statistics

The growth of nominal household incomes and a marked weakening of inflationary pressure in recent months have both helped to increase the purchasing power of households. However, the resumption of income growth has been weak in contrast to developments in the housing market, where house prices have been pushed up by growing demand and the continued relatively low level of housebuilding. Average house prices in England and Wales increased by 10.3% in 2014 – up from 3.8% in the previous year – outstripping the growth of household incomes. This has in turn raised questions about affordability and household exposure to debt. While the aggregate stock of long term borrowing held by households has fallen since the onset of the economic downturn, the pace of that decline has slowed in recent quarters as the housing market has started to recover.

While these national level trends are significant, the housing market varies substantially across the different regions of the UK. Closer examination of house prices and income at the local level can provide a more detailed assessment of trends in housing sustainability and affordability. ONS have recently published the latest ‘House Price Statistics for Small Areas’ (HPSSA) data, that show the median sale price of all dwelling types (not mix-adjusted) sold in 348 local authorities across England and Wales. This price information helps provide a more detailed local picture of housing affordability.

Figure 7 shows the distribution of annual house price growth across local authorities. The percentiles of the distribution are shown, indicating the percentage of local authorities which experienced house price growth at or below the stated rate. Half of all areas, for example, experienced house price growth above the 50th percentile rate – also known as the median rate. Growth in the headline average house price index (HPI) for England and Wales is shown for comparison.

Figure 7: Annual mix adjusted house price growth in England and Wales, and the distribution of house price growth in the HPSSA, %

Source: Office for National Statistics

In general, annual average HPI growth for England and Wales follows a similar trend to median house price growth across the local authorities. The England and Wales HPI grew at an average annual rate of 7.4% between 1996 and 2014, whereas the 50th percentile or median growth rate averaged 6.8%. However, the breadth of experiences shown in Figure 7 is striking. The spread of house price growth peaked around 2003-04, and although it narrowed over the following ten years, the difference remains substantial. In 2009, for instance, although many areas saw house prices fall, 10% of areas saw house prices rise by at least 0.4%. Equally, despite headline house price inflation of 7.9% in 2010, at least 10% of all areas saw median house prices fall over this period.

The localised nature of the housing market means that the experience of house price growth for any given local authority may differ substantially from the headline rate. Table 1 examines how these diverse experiences affected the distribution of house price levels between 1999 and 2014. On this measure, the lowest-priced 10% of local authorities had median house prices of less than £46,000 in 1999, rising to £120,000 in 2014. By contrast, the threshold for an area to be among the highest-priced 10% of local authorities increased from around £121,000 to £335,000 over the same period.

The pace of the recovery of house prices after the economic downturn has also varied. The 90th percentile – the median price threshold that 10% of local authorities exceed in a given year – recovered to its pre-downturn peak fastest, in just one year. Since then the 90th percentile has grown at a consistently faster pace than other percentiles, as shown in Table 1. This reflects faster house price growth in areas where houses are relatively expensive. The 10th, 30th, 50th, and 70th percentiles have also grown since 2009, albeit at a slower pace.

One important contribution to this recent divergence in median house prices comes from a regional effect. London and the South East experienced the highest house price inflation in England and Wales in 2014, growing by 17.4% and 9.9% respectively. Mix-adjusted average house prices for 2014 show London and the South East as the only regions valued above £300,000, with average prices of £490,000 and £320,000 respectively.

Variation in house prices is matched by substantial variations in income, which in turn results in differences in housing affordability across the UK. For example, the 2014 Annual Survey of Hours and Earnings (ASHE) showed that the median level of earnings at the local authority level ranged from £16,126 (in Blackpool) to £36,519 (in Westminster). There was also a wide range of growth in local authority median earnings compared to the previous year, from a fall of 10.5% (in Bolsover) to growth of 15.2% (in Tandridge).

Figure 8 combines median house prices with median individual earnings to summarise one measure of housing affordability. For selected years, it shows the distribution of the median house price to median earnings ratio among local authorities. It shows that in 2007, 118 local authorities had a median house price to median individual earnings ratio of between 7 and 9 – a slightly higher peak than in the other years shown. Note that because the earnings data used here is earnings data for individuals, the ratios are substantially higher than those for households, which frequently combine different sources of income for multiple individuals.

Figure 8: Number of local authorities by house-price-to-earnings ratio

Source: Office for National Statistics

Notes:

1. Due to some of the data being disclosive, ratios have been obtained for 307 of the 348 local authority areas.

Figure 8 suggests that housing affordability has varied both across local authorities and through time. Between 2002 and 2007, the distribution of median house price to median earnings ratios shifted to the right, indicating an increase in the house price earnings ratio across many local authorities. The median ratio increased from 6.4 in 2002 to 8.8 in 2007, reflecting stronger house price growth than earnings growth across many local authorities, as households became increasingly leveraged, taking on larger quantities of mortgage debt.

Following the economic downturn, however, this trend reversed as house price growth slowed markedly faster than wage growth. In 2013 and 2014, the distribution of house price to earnings ratios is shifted to the left relative to the profile seen in 2007. However, in 2014 the distribution also widened and became more positively skewed, with a higher frequency of local authorities experiencing ratios of between 14 and 16 in particular. The majority of these regions with high ratios were located in London or the South East, with the remainder located in the South West and East of England.

To unpack these most recent developments in more detail, Figure 9 divides the distribution of house price to earnings ratios for 2014 into the contributions of four types of local authority. It divides areas into four groups by comparing median house prices and earnings at the local authority level to the median local authority values of these respective series. It shows that most of the lowest house price to earnings ratios appear among local authorities where house prices and earnings are both relatively low. By contrast, the top of the distribution is largely composed of local authorities with both high house prices and high earnings: the remaining areas in this group tend to have below median incomes – suggesting that affordability in some areas is a particular issue.

Figure 9: Number of local authorities by house-price-to-earnings ratio, broken down by the median house price and earnings characteristics, 2014

subtitle

Source: Office for National Statistics

Finally, while this analysis is broadly indicative, it has several important limitations. Firstly, the data are based on mix-unadjusted house prices so may reflect the changing mix of houses that are sold in a given year. It uses median individual earnings in each area and does not consider other sources of household income. Finally, note that simply because house price to earnings ratios are high does not mean that individual household indebtedness is high: this depends on the proportion of individuals moving each period, household wealth and the means of payment.

Table notes:

1. Percentage change on previous period, seasonally adjusted, CVM
2. Levels, seasonally adjusted, CP
3. Expressed in £ billion
4. Percentage change on previous period, seasonally adjusted, CP
5. Public Sector net borrowing, excluding public sector banks. Level change on previous period a year ago, not seasonally adjusted

Table notes:

1. Monthly data shows a three month rolling average (e.g. The figure for February is for the three months Jan - Mar)
2. Headline employment figure is the number of people aged 16-64 in employment divided by the total population 16-64
3. Headline unemployment figure is the number of unemployed people (aged 16+) divided by the economically active population (aged 16+)
4. Headline inactivity figure is the number of economically active people aged 16 to 64 divided by the 16-64 population
5. Percentage change on previous period a year ago, seasonally adjusted
6. Estimates of total pay include bonuses but exclude arrears of pay (£)
7. Calculated by JSA claimants divided by claimant count plus workforce jobs
8. Percentage change on previous period a year ago, non-seasonally adjusted